<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Breakout Brainstorming Session</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Group</string-name>
        </contrib>
      </contrib-group>
      <pub-date>
        <year>2014</year>
      </pub-date>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>Breakout</p>
      <p>Group 1</p>
      <p>Longevity of URIs in a changing environment.</p>
      <p>Changes of vocabularies/datasets while keeping the same
namespace can cause applications to fail.</p>
      <p>How do we (and particularly industry!) know which
vocabularies and technologies that have strong community
backing?</p>
      <p>Longevity of URIs in a Changing Environment
Newly developed vocabularies believed to reach external
uptake should be hosted by W3C, or possibly PURL.org.
For existing vocabularies; log and publish statistics on read
access so that it can be determined which existing
vocabularies are heavily used, in order to learn how
signi cant this potential problem is.</p>
      <p>Possibly consider how BitTorrent anchor links work; could
their approach to DNS-less resource resolution work also
for Semantic Web data?</p>
      <p>Changes to Vocabularies</p>
      <p>Never change a published (i.e., available via a publicly
resolvable URI) vocabulary or dataset - only add new
URI:s with new versioning info.</p>
      <p>Further work needed on Ontology Evolution and logging,
particularly regarding presenting change deltas in an
accessible user-friendly or at least developer-friendly way.
Related to the above: the use of the deprecation warning
and versioning annotations in OWL needs to be increased.
Could a system be developed to generate reports from a
codebase, detailing which datasets/vocabularies/libraries
that the code depends upon?
Emphasize the use of and development of indices like LOV
and LodStats, possibly also for other types of artefacts
than just vocabularies.</p>
      <p>Could a social networking-based community portal be of
use? Currently the semantic web research community
exists across a great deal of sites, keeping it all in one place
and combining this with a user-maintained directory of
quality components and technologies might be bene cial.</p>
    </sec>
  </body>
  <back>
    <ref-list />
  </back>
</article>